Fig. 6: scRNA-seq analysis validates key cell-type identities constituting the human SEM.

a, UMAP of single cells from human SEMs coloured by the cell clusters and identified on the basis of known marker genes and gene signatures (Methods). Cluster sizes: 1,963 (cluster 0); 1,483 (cluster 1); 1,431 (cluster 2); 1,344 (cluster 3); 1,265 (cluster 4); 957 (cluster 5); 905 (cluster 6); 898 (cluster 7); 662 (cluster 8); 448 (cluster 9); 441 (cluster 10); 265 (cluster 11); and 128 (cluster 12). b, Normalized expression of key markers projected on the UMAP. Cell-type clusters are highlighted in red. c, Dot plot illustrating the expression of key markers across the 13 cell type clusters. Subcluster 10* indicates a subgroup of cells (n = 31) from cluster 10 that were re-annotated as CTb (cytotrophoblast-like cells) (Methods and Extended Data Figs. 12d and 14). The colour intensity indicates average expression. The dot size indicates the percentage of cells in the cluster that express the marker. Com., committed; Post. posterior. d, Top, heatmap showing normalized expression of PGC markers (SOX17, PRDM1, NANOS3, NANOG and OCT4). Nine cells were OCT4⁺PRDM1⁺SOX17⁺ and NANOS3⁺. Bottom, OCT4⁺PRDM1⁺SOX17⁺ PGC-like cells (n = 27) were identified in epiblast-like clusters, with the majority in cluster 4 (n = 19, green). e, Left, UMAP of integrated human embryonic reference data²⁷ consisting of six human embryonic datasets spanning early zygotes, in vitro cultured human blastocysts^29,30,31, 3D in vitro cultured human blastocysts until pre-gastrulation stages³², and a CS7, 16–19 d.p.f., human gastrula³³. The colour of each data point corresponds to the cell annotations from the respective publication. Right, grey data points represent embryonic reference cells, as in the left. Coloured triangles represent the projection positions of the neighbourhood nodes from SEM cells onto the human embryonic reference. SEM-CTb*, neighbourhood nodes representing subcluster 10* cells projected on the reference CTb.